Ceruloplasmin (CP) is a serum ferroxidase that oxidizes ferrous iron (Fe²⁺) to ferric iron (Fe³⁺) (1, 2). This oxidation promotes iron efflux from reticuloendothelial storage cells and iron incorporation into transferrin. It has been speculated that a deficiency in CP leads to a disruption in the systemic iron cycle and results in excessive iron deposition in reticuloendothelial cells (1, 2). The clinical and pathological findings in aceruloplasminemia strongly support this hypothesis (3–6).

Aceruloplasminemia is a disorder of iron metabolism that is characterized by excessive iron deposition in various organs such as the brain, liver and pancreas (7). Clinical manifestations of aceruloplasminemia include diabetes mellitus, retinal degeneration and neurological abnormalities (3–6). The precise pathogenesis of aceruloplasminemia is not yet clear; enhanced oxidative stress as a result of iron overload is believed to contribute to tissue injury (8–10). The disease is inherited in an autosomal recessive fashion and is caused by mutations in the ceruloplasmin (CP) gene (3–6). Since the first reports of mutations in the CP gene in three independent Japanese families with aceruloplasminemia (11–13), more than 20 such families have been investigated worldwide (5, 6).

We herein report the case of a young diabetic man with a novel splicing mutation in the CP gene. He had been followed as type 1b diabetes for the past ten years, but had not experienced any signs or symptoms indicative of brain diseases. Brain imaging studies, which were performed in order to evaluate an episode of convulsion and unconsciousness, led us to a diagnosis of aceruloplasminemia.

The patient (Fig. 1, II-5) was a 27-year-old man who, at age 17, was first admitted to our hospital because of general malaise and thirst. Laboratory tests at that time revealed that he was in a state of ketosis. His plasma glucose concentration, HbA_1c, and 24-hour C-peptide urinary excretion rate were 324 mg/dl, 12.6% (normal, 4.0–6.0%), and 4.5 μg/day (normal, 20.5–198 μg/day), respectively. He was diagnosed with type 1 diabetes and insulin therapy was initiated. Since that time, his level of diabetic control has been poor, in spite of relatively good insulin compliance and regular medical checkups. In fact, he has periodically slipped into a diabetic coma with severe ketoacidosis.

On March 26, 2002, he suddenly developed a convulsion with unconsciousness. By the time he arrived at our hospital, his convulsion had ceased. His body temperature was 35.3ºC, he had a pulse rate of 97/min, and his blood pressure was 148/98 mmHg. Laboratory tests showed that his plasma glucose level was 33 mg/dl and his HbA_1c was 8.5% (Table 1). The patient continued to manifest a disturbance in consciousness (Glasgow coma scale, E1V1M4) despite prompt normalization of his plasma glucose levels as a result of intravenous glucose administration. His unconsciousness lasted for about 15 hours after admission. We therefore conducted an extensive screening of the patient to determine the underlying cause of his condition.

After he became fully alert, his intelligence and mental status were normal. However, neurological examination revealed peripheral sensory impairment and bilaterally weak tendon reflexes in the extremities, suggesting the presence of diabetic neuropathy. Cerebellar ataxia and extrapyramidal signs were absent. Ophthalmologic examination revealed non-proliferative diabetic retinopathy, but no retinal degeneration or Kayser-Fleischer rings in the cornea. Brain MRI showed bilateral abnormal hypo-intensities in the basal ganglia, thalamus and dentate nucleus in T2-weighted images (Fig. 2A). Slight hypo-intensity was also detected only in the thalamus in T1-weighted images (Fig. 2B). Electroencephalography showed no evidence of epilepsy.

Laboratory findings showed that his serum CP (≦2 mg/dl) was below detection levels (normal, 21–37 mg/dl), his serum iron concentration was 22 μg/dl (normal, 68–174 μg/dl) and his serum ferritin concentration was 1,395.9 ng/dl (normal, 26–240 ng/dl) (Fig. 1). His serum copper concentration was 8 μg/dl (normal, 68–128 μg/dl) and his 24-hour urinary copper excretion was 26.2 μg/day (normal, 4.2–33.0 μg/day). A complete blood count revealed mild microcytic and hypochromic anemia. The patient's serum ferroxidase activity, as measured by the method of Erel (14), was severely depressed (78.9 U/l, normal mean±SD of ten age-matched control subjects was 746.0±70.0 U/l).

The liver appeared to be of higher than normal density on CT scanning and showed hypo-intense signals on T1- and T2-weighted MR images (Fig. 3A). Similarly, the pancreas appeared to be hypo-intense and slightly atrophic in the MR images (Fig. 3B). Liver biopsy revealed an absence of fibrosis and cirrhosis. Berlin blue staining for the presence of iron revealed many distinctive granules in hepatocytes (Fig. 3C), but staining for copper using rubeanic acid did not uncover any abnormal findings (data not shown).

Our patient's family pedigree and laboratory findings are shown in Fig. 1. His parents were first cousins and his father (I-1) had died of pancreatic cancer at the age of 56 years. His eldest brother (II-3) had suffered from type 1 diabetes, which he developed when he was in his teens, and his laboratory findings were very similar to those of the proband (II-5). The serum CP levels in our patient's mother (I-2) and another brother (II-4) were about half the normal values.

In glucagon loading test, his fasting serum C-peptide concentration was 0.11 ng/ml (normal, 1.30±0.48 ng/ml), and his maximal serum C-peptide release six minutes after a glucagon injection (1 mg) was 0.14 ng/ml (normal, 4.73±1.14 ng/ml). Similarly, his urinary C-peptide excretion was below 0.8 μg/day (normal, 20.5–198 μg/day). In contrast, his fasting serum glucagon concentration was normal (73 pg/ml; normal, 40–180 pg/ml). Pancreatic exocrine function testing using N-benzoyl-L-tyrosyl-p-aminobenzoic acid (BT-PABA; PFD Oral, Eizai Co, Tokyo, Japan) showed a significant reduction of urinary excretion of PABA (40%; normal, 73.4–90.4%), indicating severe impairment of pancreatic exocrine function. Islet cell autoantibodies directed against glutamic acid decarboxylase and tyrosine phosphatase-like protein were not detected. Furthermore, HLA typing showed that this patient carried a haplotype that was associated with resistance to type 1 diabetes in the Japanese population (DRB1-DQB1,*1502–*0601).

After obtaining informed consent, molecular analysis of the CP gene was performed. Genomic DNA was isolated from peripheral white blood cells, and 19 exons of the CP gene and their exon-intron boundaries were independently amplified and sequenced. PCR primers that were newly designed for exons 2–12, as well as primers that were previously described (15) for exons 1 and 13–19 were used. Sequencing of amplified genomic DNA revealed a g→a transition (nt position 607+1) in the consensus splice-donor site in the intron 3 (Fig. 4A). This mutation eliminated the Rsa I restriction site in the genomic DNA. Rsa I restriction fragment length polymorphism revealed that the patient was homozygous for this mutation, and that his mother was heterozygous (Fig. 4B). His elder brother (II-3, Fig. 1), who was diagnosed with type 1 diabetes, was also homozygous for this mutation.

To determine the transcripts that were generated from this splice-site mutation, CP cDNA from our patient was analyzed. Total RNA was isolated from our patient's cultured lymphoblasts with RNAqueous™-4PCR (Ambion) and was reverse-transcribed using a Stratascript™ RT-PCR kit (Stratagene). PCR amplification was first carried out with the CP gene-specific exonic primers hCp1 (sense) and hCp16 (antisense) (12), and then the nested PCR was carried out with hCp3 (sense) and hCp8, hCp12, or hCp14 (antisense) (12). Each set of primers generated a single product in the nested PCR. Sequencing of the nested PCR products showed that the exon 3 was completely spliced out and that the exon 2 was directly linked to the exon 4 (Fig. 4C). Thus, this mutation would be expected to result exclusively in an in-frame deletion in CP mRNA, generating a mutant CP protein consisting of 975 amino acids, instead of the full protein containing 1,046 amino acids.

In this report, we described a young patient with aceruloplasminemia who presented with uncontrolled diabetes. Despite displaying due diligence in controlling his diabetes, he still experienced several episodes of hypoglycemia, as well as hyperglycemia with ketoacidosis. Glucagon loading test indicated that his capacity to secrete insulin, as expected, was severely impaired. Insulin resistance was not the case with our patient because he sometimes experienced hypoglycemic episodes. It has been reported that pancreatic insulin-producing cells were markedly reduced in number at autopsy in a patient with aceruloplasminemia (16), though it remains unclear whether this is due to a cytotoxicity induced by oxidative stress or a congenital disposition, or both (16). Taken together, we speculate that a deterioration of the function of pancreatic β cells is most likely the main cause of diabetes in our patient.

Our patient was diagnosed with aceruloplasminemia 10 years after having developed diabetes. Our first clue to his condition was his development of a convulsion and unconsciousness. While this episode was probably triggered by hypoglycemia, the unusual finding was that his unconsciousness persisted long after his hypoglycemia was corrected. It is not clear whether the presence of aceruloplasminemia accounts for this clinical finding. The brain MRI indicated an iron overload, but the patient did not show any obvious signs or symptoms of brain dysfunction when he was alert. As observed in our patient, diabetes associated with aceruloplasminemia is often difficult to control, particularly in the advanced stage. In addition, diabetes or hyperglycemia, as well as iron, is considered as an inducer of oxidative stress (17, 18). Thus, more studies will be needed so that we can come to understand the involvement of diabetes in brain dysfunction in aceruloplasminemia.

To date, more than 20 disease-causing mutations have been identified in the CP gene (5, 6). Most of these are nonsense and frame-shift mutations that cause a premature termination of translation. Exon skipping in the CP gene has been first described in our patient. RT-PCR analysis of our patient's lymphoblasts indicated that mutant CP mRNA that skipped exon 3 would be preferentially transcribed. Exon 3 encodes the latter half of domain 1 of the CP molecule, which includes His161 and His163 indispensable for the binding of copper to its trinuclear center (19). Therefore, the mutant CP produced by this patient's hepatocytes would be structurally unstable and quickly degraded. In fact, western blotting with anti-CP antibody (rabbit polyclonal anti-human CP antibody, DAKO) did not detect either a full-length protein or proteins of smaller-size in this patient's serum (data not shown). Based on the presence of iron overload as indicated by the brain MRI, it is likely that the splicing defect observed in lymphoblasts also occurs in the brain, though this needs to be investigated further.

To our knowledge, our patient is the youngest reported case of aceruloplasminemia that has been confirmed by genetic analysis. Diabetes generally precedes the development of neurological disorders in patients with aceruloplasminemia (4–6). About 65% of patients develop diabetes before 40 years of age, while the majority of patients (94%) develop neurological symptoms between the age of 40 and 60 (6). Such a delay between the onset of diabetes and neurological disorders probably accounts for the fact that a diagnosis of aceruloplasminemia is often overlooked. Diabetes associated with aceruloplasminemia rarely develops in the teens (4–6), but, as evidenced by our patient, can become manifest. Thus, aceruloplasminemia should be kept in mind as a differential diagnosis in patients with juvenile-onset diabetes.